Quantum LTO-2 tape drive and 29160N: bad wide negotiation

My system is NetBSD/i386 4.0_BETA from around May 1.

I have been using a DDS3 on a 2940A for a very long time
with absolutely
no problems (well, other than occasional media errors and
that the tapes
are too small).

I have a shiny new Quantum LTO-2 tape drive, with a wide
connector
(68-pin micro).  I got an Adaptec 29160N "Ultra
160" controller, with an
external 50-pin micro connector, because the 29160 was
64-bit PCI and
not listed as supported in NetBSD.

I have a 50-pin to 68-pin cable connecting the controller to
the tape
drive.  For reasons I don't understand the controller
negotiates 16-bit
transfers with the tape drive.

ahc2 at pci1 dev 2 function 0: Adaptec 29160N Ultra160 SCSI
adapter
ahc2: interrupting at ioapic0 pin 18 (irq 5)
ahc2: aic7892: Ultra160 Wide Channel A, SCSI Id=7, 32/253
SCBs
scsibus1 at ahc2: 16 targets, 8 luns per target

st0 at scsibus1 target 5 lun 0: <CERTANCE, ULTRIUM 2,
1914> tape removable
st0: drive empty
st0: sync (50.00ns offset 127), 16-bit (40.000MB/s)
transfers
ahc2:parity error detected in Data-in phase. SEQADDR(0x1ae)
SCSIRATE(0x95)
ahc2:parity error detected in Data-in phase. SEQADDR(0x1ae)
SCSIRATE(0x95)
ahc2:parity error detected in Data-in phase. SEQADDR(0x1ae)
SCSIRATE(0x95)
ahc2:parity error detected in Data-in phase. SEQADDR(0x1ae)
SCSIRATE(0x95)
ahc2:parity error detected in Data-in phase. SEQADDR(0x1ae)
SCSIRATE(0x95)
ahc2:parity error detected in Data-in phase. SEQADDR(0x1ae)
SCSIRATE(0x95)


I then hooked up the drive to the old controller, and it
seems to work,
but it stops/starts which isn't good.  I don't understand
why this isn't
going for 20 MB/s transfers.

ahc1 at pci1 dev 1 function 0: Adaptec 2940A Ultra SCSI
adapter
ahc1: interrupting at ioapic0 pin 22 (irq 10)
ahc1: aic7860: Ultra Single Channel A, SCSI Id=7, 3/253
SCBs
scsibus0 at ahc1: 8 targets, 8 luns per target

st0 at scsibus0 target 5 lun 0: <CERTANCE, ULTRIUM 2,
1914> tape removable
st0: drive empty
st0: sync (100.00ns offset 15), 8-bit (10.000MB/s)
transfers
st0 detached

I have 2940U2W lying around and may try that.

I will try the SCSI bios to try to set things to the right
rates.  Does
anyone have any clues?  Does anyone know if the 29160 (no N)
works?

On Mon, Jun 25, 2007 at 02:47:08PM -0400, Greg Troxel
wrote:
> 
> My system is NetBSD/i386 4.0_BETA from around May 1.
> 
> I have been using a DDS3 on a 2940A for a very long
time with absolutely
> no problems (well, other than occasional media errors
and that the tapes
> are too small).
> 
> I have a shiny new Quantum LTO-2 tape drive, with a
wide connector
> (68-pin micro).  I got an Adaptec 29160N "Ultra
160" controller, with an
> external 50-pin micro connector, because the 29160 was
64-bit PCI and
> not listed as supported in NetBSD.
> 
> I have a 50-pin to 68-pin cable connecting the
controller to the tape
> drive.  For reasons I don't understand the controller
negotiates 16-bit
> transfers with the tape drive.
> 
> ahc2 at pci1 dev 2 function 0: Adaptec 29160N Ultra160
SCSI adapter
> ahc2: interrupting at ioapic0 pin 18 (irq 5)
> ahc2: aic7892: Ultra160 Wide Channel A, SCSI Id=7,
32/253 SCBs
> scsibus1 at ahc2: 16 targets, 8 luns per target
> 
> st0 at scsibus1 target 5 lun 0: <CERTANCE, ULTRIUM
2, 1914> tape removable
> st0: drive empty
> st0: sync (50.00ns offset 127), 16-bit (40.000MB/s)
transfers
> ahc2:parity error detected in Data-in phase.
SEQADDR(0x1ae) SCSIRATE(0x95)
> ahc2:parity error detected in Data-in phase.
SEQADDR(0x1ae) SCSIRATE(0x95)
> ahc2:parity error detected in Data-in phase.
SEQADDR(0x1ae) SCSIRATE(0x95)
> ahc2:parity error detected in Data-in phase.
SEQADDR(0x1ae) SCSIRATE(0x95)
> ahc2:parity error detected in Data-in phase.
SEQADDR(0x1ae) SCSIRATE(0x95)
> ahc2:parity error detected in Data-in phase.
SEQADDR(0x1ae) SCSIRATE(0x95)
> 
> 
> I then hooked up the drive to the old controller, and
it seems to work,
> but it stops/starts which isn't good.  I don't
understand why this isn't
> going for 20 MB/s transfers.

Well, you have a 16bit-capable adapter and 16bit-capable
target, so they
negotiate 16 bits. They have to way knowing that the cable
is only 8 bits
(other than the parity error when trying to use the 16-bit
path).

> 
> I will try the SCSI bios to try to set things to the
right rates.  Does

This may help

> anyone have any clues?  Does anyone know if the 29160
(no N) works?

Yes, it does.

-- 
Manuel Bouyer <bouyerantioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la
difference
--